Environ. Sci. Technol. 2004, 38, 5010-5021
Methylmercury and Total Mercury in Plant Litter Decomposing in Upland Forests and Flooded Landscapes B R I T T D . H A L L * ,† A N D VINCENT L. ST. LOUIS Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6E 2E9, Canada
The overall objectives of this study were to examine the effects of flooding on the decomposition and mercury (Hg) content of tissues from plants common to boreal upland forests at the Experimental Lakes Area in northwestern Ontario. We used litterbags to study changes in total Hg (THg), methyl Hg (MeHg), carbon (C), and nitrogen (N) in 12 different plant tissues (birch, alder, blueberry, and Labrador tea leaves, bunchberry plants, jack pine needles, Sphagnum spp., Polytrichum spp., and Pleurozium spp. bryophytes, lichen, and fresh and extensively decomposed wood) placed on unflooded boreal forest soils and in experimentally created reservoirs over an ∼800 day period. Rates of decomposition (as indicated by differences in the percentage of C and N mass left in the tissues over time) were slower in plant tissues placed on unflooded soils compared to the same tissues that were inundated in reservoirs. Depending on tissue type and initial THg concentrations, decomposing litter on both unflooded and flooded soils was either a source or a sink for THg. Tissues where initial THg concentrations were greater than 30 ng g-1 represented a source of THg to the surrounding environment, whereas tissues that had initial concentrations of less than 30 ng g-1 gained THg mass. Initial rates of change in THg were more rapid in plant tissues placed in reservoirs compared to the same plant tissue placed on unflooded soils, but there were no differences in final THg masses after ∼800 days. Plant tissues placed in reservoirs exhibited large increases in MeHg mass, whereas MeHg mass decreased in the same plants placed on unflooded soils. This is the first study examining THg and MeHg cycling in decomposing plants in upland boreal forests and reservoirs.
Introduction Mercury (Hg) is a highly volatile metal that is easily transported from anthropogenic sources to remote areas (1). Contamination of aquatic ecosystems with Hg creates health concerns because consumption of fish is the primary means by which humans are exposed to the neurotoxic, methylated form of Hg (methyl Hg; MeHg (2)). Therefore, a solid understanding of how Hg moves from the atmosphere through watersheds is crucial to understanding the contamination of fisheries. * Corresponding author phone: (608)262-3979; fax: (608)2620454; e-mail:
[email protected]. † Current address: Environmental Chemistry and Technology Program, University of Wisconsin-Madison, Madison, WI 53706-1484. 5010
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004
Oxidized inorganic Hg (HgII) can enter watersheds in a number of ways. HgII is water-soluble and is transferred from the atmosphere to watersheds directly in wet precipitation (3). Reactive gaseous and/or particulate HgII may also be deposited to the forest canopy as dry deposition and washed off during precipitation events as throughfall (4). In the boreal ecoregion of Canada, litterfall inputs of HgII to watersheds were twice as high as HgII inputs in wet deposition in open regions (5). Although it is not fully understood how the HgII in foliage gets there, it is currently assumed that a large portion of it originates through stomatal uptake of atmospheric elemental Hg0 (6, 7). During subsequent senescence and litter decomposition, HgII may be released (8), which can in turn be transferred through watersheds in runoff to aquatic systems (9-11). Once in anaerobic zones of wetlands (12, 13), lake sediments (14), or even saturated upland soils (15, 16), HgII can be methylated to MeHg, primarily by sulfate reducing bacteria (17). MeHg then bioaccumulates through aquatic food webs (18). Although wet deposition can be an important source of HgII to watersheds far from anthropogenic sources, atmospheric deposition of MeHg to our study site (the Experimental Lakes Area (ELA) in northwestern Ontario) is relatively insignificant (3, 5). However, boreal plants contain MeHg, which is possibly obtained through uptake from soil waters (7, 19) or by methylation of HgII inside plant tissues (20). Litterfall can therefore be an important MeHg input to forest floors following senescence (5, 19). However, in well-drained upland soils, the deposited MeHg may be destroyed by microbial demethylation and evaded back to the atmosphere as reduced Hg0. Generally, upland forest soils act as sinks for atmospheric inputs of total Hg (THg; all forms of Hg) and MeHg (13) because the majority of THg and MeHg binds to organic and mineral soil particles. However, when terrestrial and wetland areas are inundated in the creation of reservoirs for hydroelectricity production or other reasons, the decomposition of flooded organic carbon in anoxic zones fuels HgII methylation, and the resulting flooded landscape becomes a source of both MeHg and HgII to reservoirs and downstream environments (21-24). Previous studies at the ELA have shown that increased rates of bioaccumulation of MeHg in fish in reservoirs results from increased production of MeHg stimulated by the decomposition of organic matter in reservoirs, not from the leaching of MeHg already stored in the decomposing organic matter (8, 24). One of the objectives of this study was to examine the leaching of THg from, and the production of MeHg in, decomposing tissues from plants common to boreal upland forests at the ELA. We used litterbags to study THg, MeHg, carbon (C), and nitrogen (N) cycling in 12 different plant tissues placed in three unperturbed boreal forest sites over an ∼800 day period. An equally important focus of this study was to explore the contribution of different decomposing plant tissues to THg leaching and MeHg production in reservoirs. This study was part of the FLooded Upland Dynamics EXperiment (FLUDEX), a whole ecosystem manipulation designed to examine the biogeochemical cycling of Hg and C in three reservoirs created over upland boreal forests that varied in amounts of organic C stored in soils and vegetation (25, 26). Because litterfall represents such an important flux of Hg to watersheds, understanding THg leaching and MeHg production in decomposing litter on soils and in reservoirs will help identify sources of MeHg contamination of fish in lakes. 10.1021/es049800q CCC: $27.50
2004 American Chemical Society Published on Web 08/21/2004
TABLE 1. Types of Plant Tissues Placed in Litterbags on Unflooded Soils and Reservoirs unflooded forests and reservoirs plant tissue living trees birch leaves (Betula papyrifera) pine needles (Pinus banksiana) wood blocks (Abies spp.) herbs and shrubs alder leaves (Alnus crispa) blueberry leaves (Vaccinium spp.) bunchberry plants (Cornus canadensis) Labrador tea leaves (Ledum groenlandicum) bryophytes Sphagnum spp. Polytrichum spp. Pleurozium spp. lichen (Cladina spp.) old wood
high C medium C low C X X X
X X X
X X X
X X X
X X X
X X X
X X X X X X
Methods Site Descriptions. The ELA consists of a number of boreal lakes set aside by the Canadian Federal and Ontario Provincial governments for limnological and ecological research. The ELA field station is located ∼50 km southeast of Kenora, Ontario, on the Precambrian Shield (27). The ELA experiences a cold temperate continental climate with mean July and January temperatures of 18.5 and -17.3 °C, respectively, from 1969 to 2001. Mean annual wet deposition for this 32-year period was 699 mm, with ∼25% falling as snow. Upland areas at the ELA ranged from open lichen-covered granite/gneiss rocks to shallow nutrient-poor acidic soils supporting jack pine (Pinus banksiana), black spruce (Picea mariana), and paper birch (Betula papyrifera) forest communities. Three experimental reservoirs (with maximum depth of 2 m and average depth of 1 m) were constructed by building dikes along low-lying contours of the sites followed by flooding with water pumped from a nearby oligotrophic lake. Prior to flooding, the reservoirs had varying amounts of organic C stored in the plants and soils. Reservoirs (called Low C, Medium C, and High C [total organic C stored ) 30 900, 34 900, and 45 900 kg C ha-1, respectively]) were flooded annually in May or June from 1999 to 2001 with low-C (dissolved organic C [DOC] concentrations ∼400µmol L-1), low-Hg water (THg and MeHg concentrations ∼1.0 and 0.03 ng L-1, respectively (25)). The reservoirs were emptied each September or October to simulate drawdown in the shallow zones of northern hydroelectric reservoirs in the winter due to increased power demand. For more details on the rational behind site selection for the reservoirs as well as reservoir construction and operation, please see refs 25 and 26. Litterbag Construction. Twelve plant tissues were used in this study (Table 1). Six plant tissues were common to all three upland sites (birch, alder [Alnus crispa], and blueberry [Vaccinium myrtilloides and V. angustifolium] leaves, bunchberry [Cornus canadensis] plants, jack pine needles, and wood). The other plants were specific to individual sites (Labrador tea leaves [Ledum groenlandicum], bryophyte mosses [Sphagnum spp., Polytrichum spp., and Pleurozium spp.], lichen [Cladina spp.], and old wood; Table 1). In the spring of 1999, birch, alder, blueberry, and Labrador tea leaves and jack pine needles were picked from trees and shrubs. Vinyl gloves were worn when handling all material, and plant tissues were stored in clean plastic Ziploc bags. Bunchberry samples consisted of both the aboveground leaves and stems. Sphagnum spp. were collected from wetland patches in
upland forests, and the capitulum (growing tip) was removed from each plant to ensure there was no regrowth in situ. Upland dwelling bryophytes (Pleurozium spp. and Polytrichum spp.) were collected from large patches of moss located in a nearby old growth forest and trimmed to exclude the rhizoids. Lichen (Cladina spp.) was collected from exposed bedrock on a ridge top and trimmed to exclude soil. All samples were dried at ambient temperature for two weeks on labmat in a clean laboratory at the ELA to avoid contamination of plant samples. Large amounts of plant tissue were retained at the beginning of the study to measure initial concentrations of C, N, THg, and MeHg. Approximately 5 g dry weight (d.w.) of each plant tissue was placed in a 10 × 10 cm litterbag constructed of acid washed 400 µm Nitex mesh sewn together using heavy nylon thread. Litterbags were also filled with wooden blocks (∼13 g) cut from a single piece of fir lumber containing no sapwood and knots and ∼5 g pieces of extensively decomposed wood (‘old wood’) collected from fallen decomposing trees in the upland sites. Litterbag Placement. To compare decomposition rates and changes in THg and MeHg content in plants on unflooded soils and in the reservoirs, five sets of triplicate litterbags containing each plant tissue type were placed in the three reservoirs and on top of existing litter in the forest immediately adjacent to each reservoir. We also wanted to compare rates of decomposition at different depths in the reservoirs, so five sets of triplicate litterbags were placed in shallow (1 m) zones of each reservoir. Our forest and reservoir sites differed in areal coverage and species composition of plant communities, so litterbags placed in each site contained plant tissues characteristic of the site. The High C site was a wet jack pinedominated forest, with an understory of wetland plants. Litterbags containing the six common plant tissues (birch, alder, and blueberry leaves, bunchberry plants, jack pine needles, and wood), Sphagnum spp. moss, and Labrador tea leaves were placed in a hollow in the wet jack pine dominated forest community outside of the flooded area and at two locations in the 0.74 ha High C reservoir. The Medium C site was a dense jack pine forest with birch and alder, with an understory of blueberry shrubs and various bryophytes and herbs. Litterbags containing the six common plant tissues as well as old wood were placed in this homogeneous forest as well as in two locations in the 0.50 ha Medium C reservoir. The Low C site had shallow soils supporting sparse stands of jack pine and birch with a blueberry shrub dominated understory and areas of thin glacial till with lichens, bryophytes, blueberry shrubs, and exposed bedrock. Litterbags containing the six common tissues as well as bryophytes (Pleurozium spp. and Polytrichum spp.) and lichen were placed in the jack pine, birch, and blueberry dominated forest stand and in two locations in the 0.63 ha Low C reservoir. Litterbag Collection and Processing. Three litterbags of each plant tissue type were retrieved from all unflooded and flooded sites in autumn 1999 (the first summer that the reservoirs were flooded) and in spring and autumn of both 2000 and 2001. Any foreign debris on the outside of each litterbag was removed by gloved hand. Intact litterbags were then allowed to dry on labmat in a clean laboratory at the ELA at ambient temperature for two weeks. Once dry, plant tissues were removed from the litterbags, weighed, and freezedried. Freeze-dried tissues from each litterbag were ground using an acid-cleaned coffee grinder, with the exception of wood blocks, which were shaved using a clean stainless steel rasp. A portion of each dried sample was analyzed for concentrations of C, N, THg, and MeHg. Analytical Methods. Carbon and Nitrogen. C and N concentration in plant tissues from each litterbag were analyzed using an Exeter Analytical Model 440 elemental analyzer at the University of Alberta Limnology Laboratory VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5011
(Edmonton, AB) and a Carlo Erba EA1108 elemental analyzer at the University of Waterloo Environmental Geochemistry Laboratory (Waterloo, ON). Concentrations of samples analyzed at both labs were always within 20% of each other. The mass of C or N in each litterbag was calculated by multiplying the C or N concentration by the final air-dried weight of tissue in the litterbag. The percent mass of C or N remaining in each litterbag at the time of sampling was standardized to the initial C or N mass placed into each litterbag (see below). C:N ratios were also calculated for each sample. Mercury. All Hg analysis was performed in the University of Alberta Low-level Hg Analytical Laboratory. For THg and MeHg analyses, equal amounts of a particular freeze-dried tissue from the three triplicate litterbags sampled from one site were pooled together and analyzed to provide one THg or MeHg concentration per tissue per site per sampling date. Due to the high cost associated with Hg analysis, THg and MeHg were only analyzed for three out of five sampling dates (each autumn sampling date). Samples from all three locations (deep, shallow, and unflooded) were analyzed for THg, whereas only the deep and unflooded samples were analyzed for MeHg (see Table S1 in Supporting Information). THg analyses were performed on 50-200 mg of ground, freeze-dried plant tissue digested in 7 mL of 7:3 (vol:vol) HNO3:H2SO4 in sealed Teflon digestion bombs. Samples were initially digested at 125 °C for 2 h, after which 1 mL of BrCl and 19 mL of distilled, deionized water were added. Samples were then heated overnight at 60 °C. 50 µL-3 mL of digested sample was placed into glass bubblers with ∼150 mL of Milli-Q water, 1 mL of SnCl2 (to reduce all HgII to Hg0), and 200 µL of NH2OH*HCl (to neutralize the oxidation to HgII). Samples were purged for 10 min using ultrahigh purity (UHP) N2. During the purging process, Hg0 was transferred to traps containing gold-coated glass beads (gold traps) and dried in the N2 stream for 10 min. Mercury on gold traps was then thermally desorbed to the analytical system, and THg was detected using cold vapor atomic fluorescence spectrophotometry (CVAFS) as described in ref 28 with detection at 0.1-0.3 ng g-1 at a blank level of 0.3-0.4 ng L-1. Standard reference material (NRC-TORT) and standards were analyzed in tandem with all samples, and concentrations were always within 10% of certified values. Spike recoveries were always within 90-110% of original spike addition. 10% of THg samples were analyzed in duplicate. For MeHg analysis, ∼300 mg of processed plant tissues was digested overnight at 60 °C in Teflon bottles containing 5 mL of 25% KOH in methanol as described in refs 29 and 30. However, some important changes were made to this method. These previously reported methods involved the addition of a subsample of the digested sample directly to MeHg bubblers, but when we did this the bubblers were overwhelmed by the large volume of KOH/methanol solution that needed to be added to detect the sometimes low concentrations of MeHg in plant tissues. Instead, a 1 mL aliquot of the digest was centrifuged, and 500-800 µL of supernatant was distilled at 135 °C in 95-100 mL of Milli-Q water, 1 mL of 50% H2SO4, 1 mL of 25% CuSO4, and 0.5 mL of 20% KCl until 85% of the sample was transferred to the receiving jar. The distillate was then placed in glass bubblers with 500 µL of acetate buffer and 100 µL of Na tetraethyl borate and purged onto Carbotraps using UHP N2. Mercury species were thermally desorbed from Carbotraps, separated using a gas chromatography column, reduced using a pyrolytic column, and detected by CVAFS (31-33) (detection limits of 0.1-0.3 ng g-1 at a blank level of 0.05 ng L-1). Spike recovery for MeHg analysis was between 80% and 120% of original added spike and concentrations in standard reference material (NRC-TORT) and analytical standards were always within 10% of certified values. 10% of MeHg samples were 5012
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004
analyzed in duplicate. As a comparison, samples were also analyzed using a direct distillation of plant tissues in 2:1:2 (vol:vol:vol) in H2SO4:KCl:CuSO4. Samples that underwent initial digestion in KOH/methanol prior to distillation resulted in higher recoveries and better replicates. As a further check, samples of ground Pleurozium spp. and bunchberry plants taken at the beginning of the study were sent to Flett Research Ltd. (Winnipeg, MB) and analyzed using similar methods to ours (initial digest of 5-10 mg of tissue in 300 µL of KOH/ methanol, followed by ethylation and analysis on CVAFS). Average MeHg concentrations in Pleurozium spp. and bunchberry plants analyzed in our lab (1.43 ( 0.41 and 0.30 ( 0.02 ng g-1, respectively) were very similar to those analyzed by Flett Research (1.60 ( 0.22 and 0.35 ( 0.02 ng g-1, respectively). The mass of THg and MeHg in each litterbag was calculated by multiplying the THg or MeHg concentration by the final air-dried weight of tissue in the litterbags, giving us three THg masses for each tissue, from each site, on each date. The percent mass of THg or MeHg remaining in each litterbag at the time of sampling was standardized to the initial THg or MeHg mass in each litterbag. Data are presented as an average for unflooded or reservoir sites ( one standard error; however, because the samples were pooled, standard errors on THg and MeHg masses reflect changes in tissue mass. Statistical Analysis. We were primarily interested in determining differences between C, N, THg, and MeHg mass in plant tissues placed on unflooded soils and those placed in reservoirs, as opposed to differences in masses among the three different forest sites or three different reservoirs. As well, detailed statistical analysis could not be performed on THg and MeHg data from individual forests and reservoirs because plant tissue was pooled prior to analysis. Therefore, we tested combined data from all three reservoirs against combined data from all three unflooded sites. We also wanted to determine differences in C and N mass in plants placed in shallow (1 m) locations. Therefore, within each sample date, we tested for differences in C and N mass among plant tissue placed in deep and shallow locations in the reservoirs and on unflooded soils. Similarly, we tested for differences in THg mass in plant tissue placed among deep and shallow locations in the reservoir and plant tissue placed on unflooded soils, for each fall sampling date. We tested for differences between MeHg mass in plants tissues placed on unflooded soils and MeHg mass in plant tissue placed in deep regions of the reservoirs. ANOVAs and multiple comparison procedures (SigmaStat Version 3.00) were used for all statistical analysis (34), and p values less than 0.05 were considered significant.
Results and Discussion Initial C and N Concentration and C:N Ratios in Plant Tissue. Initial C concentration in birch, blueberry, alder, and Labrador tea leaves, jack pine needles, wood, and old wood was between 49.8 ( 0.04% C and 56.6 ( 0.1% C (Table 2). The lowest C concentrations were found in bunchberry plants, bryophytes (Sphagnum spp., Pleurozium spp., and Polytrichum spp.), and lichens. The highest C concentration was found in the old wood. N concentration in plant tissues ranged from 0.03 ( 0.01% N in old wood to 2.8 ( 0.1% N in the leaves of alder, a known N fixing species. Ratios of C:N were lowest in birch, blueberry, and alder leaves and bunchberry plants and highest in old wood and wood (Table 2). Initial MeHg and THg Concentrations in Plant Tissue. Initial THg concentrations in plant tissue ranged from under 5 ng g-1 to over 90 ng g-1 (Table 2). THg concentrations were lowest in blueberry and birch leaves, wood, old wood, and bunchberry plants (3.6-9.8 ng g-1). Moderate THg concen-
TABLE 2. Tissue Mass, Carbon (C) and Nitrogen (N) Concentrations, C:N Ratios, and Total Mercury (THg) and Methylmercury (MeHg) Concentrations in Tissues in Litterbags Prior to Placement in Natural Forests and Reservoirs plant tissue living trees birch leaves pine needles wood blocks herbs and shrubs alder leaves blueberry leaves bunchberry plants Labrador tea leaves bryophytes Sphagnum spp. Polytrichum spp. Pleurozium spp. lichen old wood
tissue mass (g litterbag-1)
C concentration (% C)
N concentration (% N)
C:N
THg (ng g-1)
MeHg (ng g-1)
5.12 ( 0.02 5.16 ( 0.01 12.81 ( 0.04
49.8 ( 0.04 52.1 ( 0.1 52.1 ( 0.4
2.4 ( 0.05 1.3 ( 0.01 0.2 ( 0.01
21.1 41.0 226.4
7.13 14.07 3.62
0.16 0.09 0.23
5.11 ( 0.01 5.06 ( 0.004 5.10 ( 0.01 5.37 ( 0.03
52.1 ( 0.1 51.4 ( 0.1 46.1 ( 0.4 53.3 ( 0.3
2.8 ( 0.12 2.3 ( 0.02 2.4 ( 0.02 1.6 ( 0.03
18.4 22.3 19.1 33.5
12.84 5.75 9.77 27.13
0.02 0.19 0.30 0.45
5.07 ( 0.01 5.10 ( 0.01 5.01 ( 0.01 5.15 ( 0.01 5.17 ( 0.04
46.6 ( 0.3 47.5 ( 0.3 48.4 ( 0.2 45.8 ( 0.02 59.6 ( 0.1
1.0 ( 0 1.5 ( 0.07 1.0 ( 0.02 0.1 ( 0.01 0.03 ( 0.01
46.6 31.0 46.5 76.3 198.6
52.57 93.85 64.64 33.17 9.68
0.55 0.40 1.43 0.56 0.02
trations were found in Labrador tea leaves and lichens (27.1 and 33.2 ng g-1, respectively), whereas bryophytes had the highest THg concentrations (52.6-93.9 ng g-1; Table 2). Initial MeHg concentrations were lowest in alder, birch, and blueberry leaves, bunchberry plants, jack pine needles, wood, and old wood (0.02 to 0.19 ng g-1; Table 2). Intermediate MeHg concentrations were found in bunchberry plants, Polytrichum spp., Sphagnum spp., Labrador tea leaves, and lichens (0.30-0.56 ng g-1), while the feather moss, Pleurozium spp., had the highest initial MeHg concentrations (1.43 ng g-1; Table 2). Initial MeHg concentrations were similar to those found in an earlier survey of MeHg in plants at the ELA (20). Decomposition of Litter on Unflooded Soils and in Reservoirs. Decomposition of Plant Tissues in Unflooded Forests. Loss of tissue mass from all plant tissues placed on unflooded soils, with the exception of wood and old wood, was highest during the first year, but decomposition continued at gradual and constant rates in subsequent years, with final losses of between 20%-80% of the original mass after 800 days (Figures 1A and 2A). Wood and old wood exhibited very little to no tissue mass loss during the study period (Figure 3A). All plant tissues placed on unflooded soils, with the exception of Sphagnum spp., exhibited a gradual decline in C concentration, losing up to 20% of their original C over the three-year study. Average C concentration in Sphagnum spp. increased by 5% by the end of the study. By the end of the study, N concentration increased by 5% to 40% in all tissues except lichen (which decreased by 15%) and wood (which increased by 85%; see Table S2 in Supporting Information). Trends in C and N mass loss generally reflected changes in tissue mass. Birch, alder, blueberry, and Labrador tea leaves and bunchberry plants on unflooded soils exhibited gradual declines in C and N mass, with final losses as high as 80% and 70% of original C and N mass, respectively (Figures 1 and 2, B and C). C and N mass in bryophytes and lichens placed on unflooded soils also declined but at slower rates than the more labile tissues (between 10% and 20% of original C and N mass lost). Wood and old wood showed very little C mass loss but exhibited up to 30% increase in N mass (Figure 3B and C). Birch, alder, and blueberry leaves and bunchberry plants demonstrated sharp decreases in C:N ratios after the first summer but tended to stabilize after the first year (Figures 1D and 2D). C:N ratios in jack pine needles, wood, and old wood had constant declines over time and by the end of the study were generally 10%-50% lower than original values (Figures 1D and 3D). Decreases in C:N ratios also occurred
after the first season of study in bryophytes and lichens; however, these decreases were less than those observed in other tissues. Mid-way through the study, C:N ratios in bryophytes were similar to preflood values but had decreased by 20% by the end of the study (Figure 2D). Tissue decomposition rates in unflooded forest sites were similar to those found in other studies examining plant tissue decomposition in the boreal forest ecoregion (35-37). Two main factors interact to affect leaf litter decomposition rates in terrestrial ecosystems (38). The first is the micro- and macroclimates (most notably temperature and moisture conditions (39) and light (40)) to which the litter is exposed. The second factor is related to the chemical nature of the litter or litter quality (38). Globally, the effects of litter chemistry on plant decomposition are secondary to the effects of climatic conditions (39). However, locally, where decomposing plants experience similar climates, litter quality primarily determines litter decomposition. Litter quality is often defined by the C:N ratio in plant tissues. Terrestrial autotrophs generally have low N concentration, and C:N ratios in terrestrial plants range from 7.5 to 225, with a mean of 36 (41, 42). Initial C:N ratios in our plant tissues are within published ranges, with a low of 18.4 for alder leaves, a shrub capable of N2 fixation, and a high of 226 for wood. Studies have found tissues with large C:N ratios (greater than 25) experience little net mineralization of N and therefore decompose slowly over time (43). Based on initial C:N ratios in our plant tissues, we expected to see higher decomposition rates in birch, alder, and blueberry leaves and bunchberry plants, very little decomposition in wood and old wood, and intermediate rates in the other tissues. Our data supported this hypothesis. Decomposition of Plant Tissues in Reservoirs. All flooded plant tissues, with the exception of jack pine needles, wood, and old wood, had rates of initial tissue mass loss that were ∼2 times greater compared to the same tissues placed in the unflooded forest sites (Figures 1A′ and 2A′). Jack pine needles experienced similar tissue mass loss in both unflooded and flooded locations (Figure 1A′). As on unflooded soils, wood and old wood exhibited very little to no mass loss in the reservoirs during the study period (Figure 3A′). As with tissues placed on unflooded soils, average C concentration in all flooded plant tissues decreased, while N concentration increased (with the exception of bryophytes in which N concentration decreased). Changes in C and N concentration of plant tissues placed in the reservoirs were either similar to, or greater than, those placed on forest soils (see Table S2 in Supporting Information). VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5013
FIGURE 1. Changes in decomposition and mercury mass in foliage in unflooded and flooded forests. A: Percent of original tissue mass remaining. B: Percent of original C mass remaining. C: Percent of original N mass remaining. D: C/N ratios. E: Percent of original total mercury (THg) remaining. F: Percent of original methylmercury (MeHg) remaining. Hatched areas represent periods of inundation. C and N mass was not statistically different between tissues placed in deep and shallow locations in the reservoirs (p values ranged from 0.06 to 0.94 for all sampling dates), so we 5014
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004
averaged the C and N mass in tissues placed in all flooded locations. Final C and N masses in alder, birch, and blueberry leaves and bunchberry plants placed in the reservoirs were
FIGURE 2. Changes in decomposition and mercury mass in ground-covering plants in unflooded and flooded forests. A: Percent of original tissue mass remaining. B: Percent of original C mass remaining. C: Percent of original N mass remaining. D: C/N ratios. E: Percent of original total mercury (THg) remaining. F: Percent of original methylmercury (MeHg) remaining. Hatched areas represent periods of inundation. significantly lower than those placed on unflooded soils (p values between deciduous leaves (birch, alder, blueberry, Labrador tea) > jack pine needles > bryophytes and lichens > wood and old wood. However, environmental factors such as temperature and dissolved nutrient concentrations are also important (43, 46). Generally, litter decomposition is accelerated in warmer waters that have higher concentrations of dissolved N and P (47). There were no differences in changes in C and N mass among reservoirs, which suggests that environmental factors affecting decomposition were similar in all three reservoirs. In fact, there were little differences in water column (48) and soil (26) temperatures and dissolved N and P concentrations among reservoirs within a given year (unpublished data, R. A. Bodaly and A. Majewski, Freshwater Institute, Winnipeg, MB). Changes in THg in Decomposing Litter on Unflooded Soils and in Reservoirs. THg Concentrations. Concentrations of THg in plant tissues placed in both unflooded and flooded forests changed over time (Tables 2 and 3). Generally, over the course of the experiment, THg concentrations increased in plant tissues that had initial THg concentrations less than 30 ng g-1 (birch, alder, blueberry, and Labrador tea leaves, jack pine needles, and wood). In plant tissues with initial THg concentrations exceeding 30 ng g-1 (bryophytes and lichens), THg concentrations decreased by the end of the experiment (Table 3). Changes in THg Mass. On Unflooded Soils. Overall, THg mass in birch, blueberry, alder, and Labrador tea leaves, jack pine needles, and wood increased with time (see Table S3 in Supporting Information). In unflooded forest sites the largest increases in THg mass were observed in birch and blueberry leaves and wood (Figures 1E-3E). THg mass in bunchberry plants decreased after the first season of study but increased back to initial values by the end of the study. The mass of THg in bryophytes, lichen, and old wood placed in forests decreased between 6% and 50% over time.
In Reservoirs. Because differences in C and N mass in flooded plants in deep and shallow zones were not statistically significant, THg mass in plants in deep and shallow zones in all reservoirs were averaged to obtain THg masses that were representative of all flooded locations. THg mass in flooded birch, alder, blueberry, and Labrador tea leaves, jack pine needles, wood, and old wood increased sharply after the first season of flooding (Figures 1E′-3E′). Generally, rates of change in THg mass in the first season of flooding were significantly greater in tissues in reservoirs than in the same tissues on unflooded soils (p value ) 0.009 for all tissues). However, there were no significant differences in final THg mass in tissues placed in reservoirs compared to those placed on forest soils (p values ) 0.985 and 0.341 for all tissues sampled in September 2000 and September 2001, respectively). The mass of THg also increased in flooded alder and Labrador tea leaves, jack pine needles, wood, and old wood; however, THg increases in these tissues tended to occur only gradually over time (Figures 1E′-3E′). We observed decreases in THg mass in bunchberry plants, bryophytes, and lichen placed in the upland reservoirs. Decreases in THg mass in Sphagnum spp. were supported by results from a previous study by Heyes et al. (49) that found THg mass decreased in Sphagnum fuscum placed in litterbags in pristine and impounded wetlands. However, they also observed decreases in THg mass in other plant tissues, Carex rostrata (grass) stems and Picea mariana (spruce) needles. What Factors Control Changes in THg Mass in Decomposing Plant Tissues? In both unflooded and flooded sites, THg mass increased in plant tissues with initial THg concentrations that were less than 30 ng g-1 and decreased in plant tissues with initial THg concentrations exceeding 30 ng g-1, suggesting that changes in THg mass depend on initial THg concentrations (Figure 4A). We did not sample soil concentrations in the forest sites, but concentrations of THg in organic soils in reservoirs were measured prior to flooding from cores collected using a stainless steel barrel corer lined with plastic sleeves (25). THg concentrations were much higher in organic soils (between 78 and 89 ng g-1 d.w.; unpublished data, K. R. Rolfhus, University of Wisconsin at La Crosse, La Crosse, WI) than in decomposing plants. We believe that the organic soils were a small source of THg to decomposing plants, and transfer may occur during rain events when precipitation, throughfall, and runoff can facilitate the movement of THg from soils to plants. However, if THg in plants equalized with THg in soils only, we would have expected to see the largest increases in THg concentrations in tissues in the site with the highest organic soil THg concentrations. Since site to site differences in changes in THg mass in plants were small (48), we assume that precipitation, throughfall, and runoff, in addition to organic soils, could have been sources of THg to decomposing plants. We conclude that THg concentrations in decomposing plant tissues in natural forests and reservoirs equilibrate with THg concentrations in surrounding environments and that initial THg concentrations control changes in THg mass in decomposing tissues. Differences in Changes in THg Mass on Unflooded Soils and Reservoirs. For plant tissues decomposing both on forest soils and in reservoirs, changes in THg mass were dictated by the combined effect of tissue mass loss and equilibrium with concentrations of THg in surrounding soils and water. Generally, rates of change in THg mass were greater in flooded plants than in the same species placed on unflooded soils. These results are consistent with those of Heyes et al. (49), which found greater THg decreases in Carex stems, spruce needles, and Sphagnum fuscum in impounded wetlands compared to tissues in nonimpounded wetlands. However, for plant tissues that increased in THg mass, there VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5017
TABLE 3. Average Tissue Mass (g), Total Mercury Concentration (ng THg g-1), and Methylmercury Concentration (ng MeHg g-1) in Plant Tissue in Litterbags Sampled from Unflooded Forests and Reservoirs THg concentration (ng g-1)b
tissue mass (g)a tissue day living trees birch leaves 79 days 431 days 798 days pine needles 79 days 431 days 798 days wood blocks 79 days 431 days 798 days herbs and shrubs alder leaves 79 days 431 days 798 days blueberry leaves 79 days 431 days 798 days bunchberry plants 79 days 431 days 798 days Labrador tea leaves 79 days 431 days 798 days bryophytes Sphagnum spp. 79 days 431 days 798 days Polytrichum spp. 79 days 431 days 798 days Pleurozium spp. 79 days 431 days 798 days lichen 79 days 431 days 798 days ‘old wood’ 79 days 431 days 798 days
MeHg concentration (ng g-1)c
unflooded forest
reservoir
unflooded forest
reservoir
unflooded forest
reservoir
3.7 ( 0.2 2.6 ( 0.2 2.2 ( 0.3
1.90 ( 0.06 1.60 ( 0.07 1.40 ( 0.12
13.27 ( 1.07 30.71 ( 4.01 40.91 ( 6.42
39.12 ( 3.39 49.74 ( 3.17 67.04 ( 5.49
0.74 ( 0.03 0.34 ( 0.07 0.36 ( 0.07
0.97 ( 0.07 2.77 ( 0.33 5.93 ( 0.26
4.8 ( 0.1 3.7 ( 0.2 2.8 ( 0.3
4.37 ( 0.05 3.77 ( 0.04 3.30 ( 0.03
16.09 ( 0.07 21.92 ( 1.42 35.41 ( 2.98
15.19 ( 0.42 18.23 ( 0.49 25.51 ( 0.55
0.31 ( 0.03 0.35 ( 0.04 1.20 ( 0.16
0.45 ( 0.07 0.67 ( 0.04 1.07 ( 0.17
13.0 ( 0.1 12.3 ( 0.2 11.0 ( 0.4
12.99 ( 0.38 12.27 ( 0.14 12.18 ( 0.09
3.01 ( 0.09 7.39 ( 1.21 8.42 ( 0.87
3.49 ( 0.18 5.99 ( 0.25 4.72 ( 0.16
0.17 ( 0.001 1.52 ( 0.30 5.80 ( 1.95
0.52 ( 0.11 0.26 ( 0.01 2.11 ( 0.96
3.9 ( 0.2 2.5 ( 0.1 2.4 ( 0.2
2.37 ( 0.06 1.96 ( 0.05 1.80 ( 0.08
13.03 ( 0.74 27.38 ( 3.11 33.98 ( 1.76
30.66 ( 2.34 41.83 ( 1.25 51.89 ( 1.63
0.99 ( 0.06 0.41 ( 0.08 0.62 ( 0.07
0.93 ( 0.04 2.21 ( 0.33 3.31 ( 0.41
3.7 ( 0.1 2.7 ( 0.1 2.6 ( 0.1
2.30 ( 0.05 2.00 ( 0.04 1.86 ( 0.07
10.95 ( 0.48 24.18 ( 0.72 33.29 ( 2.01
29.26 ( 2.08 41.97 ( 1.90 52.38 ( 2.91
1.31 ( 0.42 0.57 ( 0.05 0.98 ( 0.17
3.34 ( 0.52 3.49 ( 0.34 6.62 ( 0.39
2.7 ( 0.2 1.5 ( 0.2 0.9 ( 0.1
0.83 ( 0.07 0.60 ( 0.04 0.48 ( 0.05
15.57 ( 0.28 35.92 ( 2.10 49.58 ( 1.03
84.51 ( 5.37 67.23 ( 2.58 80.63 ( 8.09
1.71 ( 0.24 0.27 ( 0.06 0.23 ( 0.02
5.44 ( 1.67 5.32 ( 1.44 3.04 ( 0.40
4.4 ( 0.1 3.3 ( 0.1 2.6 ( 0.2
3.5 ( 0.1 3.0 ( 0.03 3.0 ( 0.1
33.87 30.21 66.36
41.06 ( 8.02 41.80 ( 0.34 57.25 ( 5.69
0.43 0.54 0.28
1.37 1.95 6.22
4.4 ( 0.03 4.3 ( 0.1 3.9 ( 0.5
4.7 ( 0.1 4.5 ( 0.03 4.4 ( 0.2
56.41 38.49 47.35
32.94 ( 10.38 30.34 ( 3.20 28.22 ( 3.83
0.43 0.20 0.40
1.28 1.88 1.84
4.7 ( 0.1 4.1 ( 0.2 3.7 ( 0.1
4.0 ( 0.1 3.9 ( 0.04 3.6 ( 0.1
50.16 60.86 74.77
42.53 ( 3.01 38.63 ( 4.99 44.71 ( 5.27
0.14 0.47 0.40
1.97 4.05 5.69
4.5 ( 0.01 5.1 ( 0.5 4.7 ( 0.1
4.1 ( 0.1 3.9 ( 0.03 3.7 ( 0.04
58.72 66.38 54.00
38.44 ( 2.37 39.56 ( 5.93 45.30 ( 10.19
1.88 0.60 0.92
5.61 2.33 4.57
5.2 ( 0.1 4.5 ( 0.2 5.3 ( 0.2
3.9 ( 0.1 3.2 ( 0.3 2.5 ( 0.4
26.08 29.49 24.91
29.78 ( 0.45 22.64 ( 0.03 23.95 ( 6.34
0.36 0.10 0.36
0.99 0.77 1.57
5.1 ( 0.1 5.0 ( 0.1 5.0 ( 0.1
5.3 ( 0.1 4.9 ( 0.03 5.0 ( 0.03
7.01 7.94 9.35
9.30 ( 0.99 12.54 ( 2.48 15.26 ( 0.42
0.06 0.04 0.05
3.34 1.40 4.51
a Standard errors for tissue mass were calculated from tissue mass in plants from nine litterbags sampled from unflooded or flooded sites on one date. b Standard errors for THg concentrations in birch, alder, and blueberry leaves, pine needles, wood, and bunchberry plants were calculated from six pooled samples from three unflooded or three flooded forests sampled on one date. Standard errors for THg concentrations in Labrador tea, bryophytes, and lichens in flooded forests were calculated using two samples from one reservoir sampled on the same date. Standard errors for THg concentrations in Labrador tea, bryophytes, and lichens on unflooded soils were not calculated since concentrations for these samples came from one pooled sample from one unflooded forest. c Standard errors for MeHg concentrations in birch, alder, and blueberry leaves, pine needles, wood, and bunchberry plants were calculated from one pooled sample taken from three unflooded or three flooded sites sampled on one date. Standard errors for MeHg concentrations in Labrador tea, bryophytes, and lichens in unflooded forest were not calculated because concentrations for these samples came from one pooled sample from a single unflooded and flooded site.
were few differences in final THg masses in flooded compared to unflooded tissues. This suggests that THg mass in flooded tissues reached equilibrium with reservoir water faster than unflooded tissues did with forest soils but that equilibrium was reached in both unflooded and flooded sites within the 800-day study period. For tissues that showed a decrease in THg mass (bryophytes and lichens), final THg mass in submerged tissues were less than final THg mass in tissues on the corresponding unflooded forest soils. 5018
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004
Changes in MeHg in Decomposing Litter on Unflooded Soils and in Reservoirs. MeHg Concentrations. Over the course of the study, MeHg concentrations increased in all tissues placed on unflooded soils with the exception of Labrador tea leaves, bryophytes, and lichen (Tables 2 and 3). Final MeHg concentrations in unflooded Labrador tea leaves, Sphagnum spp., Pleurozium spp., and lichen decreased slightly to values that were lower than initial MeHg concentrations. There was no difference between initial and final
FIGURE 4. A: The percent of original total mercury (% THg) mass remaining in litterbags at the end of the experiment as a function of the initial THg concentrations. The vertical line represents initial THg concentrations of 30 ng g-1. B: The percent of original methylmercury (% MeHg) mass remaining in litterbags at the end of the experiment as a function of the initial MeHg concentration. MeHg concentrations in Polytrichum spp. placed on unflooded soils. MeHg concentrations increased in all flooded plants tissues (Tables 2 and 3). The greatest increases in MeHg concentrations were observed in flooded alder leaves, bunchberry plants, and old wood which increased to values that were over 200 times initial MeHg concentrations. Significant increases in MeHg concentrations were also observed in flooded birch and blueberry leaves (37 and 35 times over initial MeHg concentrations). Only small (3 to 14 times) increases in MeHg concentrations were observed in Labrador tea leaves, jack pine needles, bryophytes, lichen, and wood. MeHg Mass. On Unflooded Soils. Only alder leaves, jack pine needles, wood, and old wood placed on unflooded soils experienced increases in MeHg mass (Figures 1F and 3F; see Table S3 in Supporting Information). MeHg decreased in birch and Labrador tea leaves, bunchberry plants, Sphagnum spp., Pleurozium spp., Polytrichum spp., and lichen to final values that were between 31% and 72% of original MeHg mass. There was no change in MeHg mass in blueberry leaves. Increases in MeHg were unexpected in plants lying on dry
soils, since Hg methylation is as anaerobic process. It is possible that placing plants in litterbags allowed the formation of anaerobic microzones, thus promoting MeHg production. Decreases in MeHg mass in plant tissues placed on unflooded soils indicate a loss of MeHg from plant tissues decomposing in unsaturated, well-drained forest soils. Decomposing plants may therefore represented an input of MeHg to forest soils, which would then presumably be available for transport to aquatic systems. MeHg concentrations in unflooded soils were 1.13, 0.20, and 0.52 ng g-2 (unpublished data, K. R. Rolfhus), which were similar to MeHg concentrations in decomposing plant tissues. It is also possible, however, that since MeHg production is suppressed and MeHg degradation enhanced in aerobic environments (50), MeHg in plant tissues on forest soils was destroyed by microbial demethylation. In Reservoirs. All types of flooded plant tissue, with the exception of lichen, exhibited large increases in MeHg mass by the end of the study (Figures 1F′-3F′). The mass of MeHg was significantly higher in the plants placed in reservoirs compared to those placed on unflooded soils (p values ) VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5019
0.002,
